In his essay Rhetoric of the Image, Roland Barthes deconstructed a Panzani advertising image and extracted the types of messages contained within it in order to illustrate the ‘rhetoric’ of the image.
With regards to the accompanying text, he breaks it down into two parts;
Anchorage – In his words “The text helps to identify purely and simply the elements of the scene and the scene itself; it is a matter of a denoted description of the image”. From this, you can gather that the text is there to essentially describe what you can see in the image, and to make the viewer interpret it in a certain way.
Relay – In his words “Here text and image stand in complimentary relationship; the words, in the same way as the images, are fragments of a more general syntagm and the unity of the message is more realized at a higher level, that of the story”. From this, he describes that the text is a separate entity to the image, and can be used to manipulate the viewer into viewing the image meaning into a way they wouldn’t have naturally done.
Rhetoric of the Image (p. 156)